Mutant and Wild Type Human α-Synucleins Assemble into Elongated Filaments with Distinct Morphologies in Vitro *

α-Synuclein is a soluble presynaptic protein which is pathologically redistributed within intracellular lesions characteristic of several neurodegenerative diseases. Here we demonstrate that wild type and two mutant forms of α-synuclein linked to familial Parkinson’s disease (Ala30 → Pro and Ala53 → Thr) self-aggregate and assemble into 10–19-nm-wide filaments with distinct morphologies under definedin vitro conditions. Immunogold labeling demonstrates that the central region of all these filaments are more robustly labeled than the N-terminal or C-terminal regions, suggesting that the latter regions are buried within the filaments. Since in vitrogenerated α-synuclein filaments resemble the major ultrastructural elements of authentic Lewy bodies that are hallmark lesions of Parkinson’s disease, we propose that self-aggregating α-synuclein is the major subunit protein of these filamentous lesions.

␣-Synuclein is a small, 140-amino acid protein characterized by acidic stretches toward the C terminus and six repetitive, degenerate amino acid sequences of the prototype KTKEGV between amino acid residues 10 and 86 (1,2). The first synuclein identified and cloned was from the electric eel Torpedo california (3). In the latter study, the authors also cloned a rat synuclein homologue that was later termed ␣-synuclein (4). ␣-Synuclein was first associated with a human neurodegenerative disease when a 35-amino acid peptide corresponding to residues 61-95 in ␣-synuclein was purified from Alzheimer's disease senile plaques (5). This peptide was referred to as the non-amyloid component of senile plaques (NAC) 1 and its precursor protein, i.e. ␣-synuclein, was designated NACP (5). The zebra finch homologue of ␣-synuclein, synelfin, was identified as a gene product potentially implicated in neuronal plasticity (6).
␣-Synuclein is predominantly a neuronal protein expressed in brain and localized in axon terminals (5)(6)(7)(8). Because ␣-synuclein is in close proximity to and is loosely associated with presynaptic vesicles, it may modulate the function of synaptic vesicles (6,7). Although ␣-synuclein has very little structure in aqueous solution (9), it associates with small unilamellar acidic phospholipid vesicles in vitro and acquires increased ␣-helicity (10). Thus, it has been proposed that ␣-synuclein may bind vesicles by forming amphipathic helices similar to apolipoproteins (6).
A link between ␣-synuclein and Parkinson's disease (PD) first surfaced when Polymeropoulos et al. (11) reported a missense mutation (Ala 53 3 Thr) in ␣-synuclein in four Italian and Greek kindreds with autosomal dominant PD. This finding was followed by the demonstration of ␣-synuclein immunoreactivity in Lewy bodies (LBs) and Lewy neurites in patients with sporadic PD and dementia with LBs (DLB) (12). Significantly, these lesions are pathological hallmarks of PD and DLB (13,14). Subsequent studies confirmed that normal, truncated, and aggregated ␣-synuclein are major components of LBs and Lewy neurites and that antibodies to ␣-synuclein are the most reliable and consistent immunological probes for detecting these lesions in situ (15)(16)(17)(18)(19). More recently, ␣-synuclein was also shown to be a prominent component of glial cell inclusions (GCIs) and neuronal cytoplasmic inclusions (NCIs) that are characteristic of multisystem atrophy (MSA) and Hallervorden-Spatz disease (20 -24). Furthermore, a second pathogenic mutation in ␣-synuclein (Ala 30 3 Pro) also was reported in another familial PD kindred (25).
Since it is unclear how ␣-synuclein, a very soluble protein, ends up in cellular inclusions, we carried out studies to determine whether wild type (WT) and/or mutants of ␣-synuclein can self-aggregate in vitro. Here, we report that the SDS solubility of wild type and mutant ␣-synucleins is reduced after incubation in aqueous solution at physiological temperature and this change in physical property was attributed to the formation of elongated filaments.

EXPERIMENTAL PROCEDURES
Expression and Purification of ␣-Synuclein-Human WT, A30P, and A53T ␣-synuclein cDNAs subcloned into the bacterial expression vector pRK172 were expressed in Escherichia coli BL21 (DE3). Bacterial pellets were resuspended in high-salt lysis buffer (0.75 M NaCl, 100 mM MES, pH 7.0, 1 mM EDTA) containing a mixture of protease inhibitors, heated to 100°C for 10 min, and centrifuged at 70,000 ϫ g for 30 min. The supernatants were dialyzed against 10 mM Tris, pH 7.5, applied to a Mono Q column, and eluted with a 0 -0.5 M NaCl gradient. Protein concentration was determined using the bicinchoninic acid protein assay (Pierce) and bovine serum albumin as a standard.
Centrifugal Sedimentation-Following incubation at 37°C in 100 * This work was supported by grants from the National Institute on Aging and a Pioneer Award from the Alzheimer's Association. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertise- mM sodium acetate, pH 7.0, with continuous shaking, samples were centrifuged at 150,000 ϫ g for 30 min, SDS-sample buffer (10 mM Tris, pH 6.8, 1 mM EDTA, 40 mM dithiothreitol, 1% SDS, 10% sucrose) was added to pellets and supernatants and heated to 100°C for 15 min. ␣-Synuclein was resolved on SDS-PAGE, stained with Coomassie Brilliant Blue R-250, and quantified by densitometry.
Antibody Generation-SNL-1 and SNL-4 are affinity-purified rabbit polyclonal antibodies raised against peptides corresponding to amino acid residues 104 -119 and 2-12 in ␣-synuclein, respectively. Syn 202, Syn 204, Syn 205 and Syn 208 are novel mouse monoclonal antibodies raised to synucleins, as described previously (20). The epitopes for Syn 202 and 205 are localized to amino acid residues 130 -140, while the Syn 204 and 208 epitopes map to residues 87-110. The relative activities of the antibodies were determined by enzyme-linked immunosorbent assay using ␣-synuclein as the antigen and by performing serial dilutions of the antibodies.
Immunoelectron Microscopy and Negative Staining-␣-Synuclein filaments were decorated with anti-␣-synuclein antibodies and negative stained with uranyl acetate as described previously (27). Briefly, assembled ␣-synuclein filaments were absorbed to 300 mesh carbon coated copper grids and stained with 1% uranyl acetate or labeled with antibodies to ␣-synuclein followed by secondary antibodies conjugated to 10 nm gold and staining with 1% uranyl acetate.

RESULTS
The purity of recombinant ␣-synuclein proteins was demonstrated in Coomassie Blue-stained SDS-PAGE gels (Fig. 1A).
No contaminating proteins were seen even when 100 g of purified recombinant ␣-synucleins were loaded in separate lanes of an SDS-PAGE gel (data not shown). Incubation of 5 mg/ml WT ␣-synuclein at 37°C for 48 h in a number of different buffers resulted in the aggregation of ␣-synuclein as reflected by decreased mobility on SDS-PAGE (Fig. 1B). Diffuse smears of the incubated proteins on SDS-PAGE gels may reflect the rapid association/dissociation of ␣-synuclein in the presence of SDS. Temperature was a major determinant of ␣-synuclein aggregation, since under the same conditions, incubation at 37°C generated abundant aggregation that was not significantly detected when the protein was incubated at room temperature (Fig. 1B). WT (Fig. 1C), A30P (Fig. 1D), and A53T (Fig. 1E) ␣-synucleins demonstrated a similar ability to aggregate, although A53T ␣-synuclein seemed to aggregate to a slightly greater extent at lower concentrations. The ability of ␣-synuclein to polymerize was confirmed with centrifugal sedimentation experiments (Fig. 2). WT, A30P, and A53T ␣-synuclein polymerization was concentration-and time-dependent, and the A53T mutant had a greater propensity to polymerize.
␣-Synuclein filaments were only modestly labeled with antibodies to the N terminus (SNL-4) (Fig. 4A) or the C-terminal region (SNL-1; Syn 202 and Syn 205) (Fig. 4, B and C; data not shown) of ␣-synuclein. However, antibodies to epitopes within the central part of ␣-synuclein (Syn 204 and Syn 208) demon- ␣-Synuclein Filament Assembly 7620 strated very strong labeling (Fig. 4, D-F). Antibodies SNL-1, SNL-4, Syn 202, and Syn 205 were used at 10 -20-fold higher relative immunoreactiveactivity than antibodies Syn 204 and Syn 208. Thus, these results may suggest that the N-and C-terminal regions of ␣-synuclein are less accessible than the central region within these filaments, and this may imply that both ends of the polypeptide are involved in polymer formation and embedded within the filaments. DISCUSSION Despite the fact that ␣-synuclein is a small soluble synaptic protein that is largely devoid of secondary structure in aqueous buffer, we show here that mutant and WT recombinant ␣-synucleins polymerize into morphologically distinct filaments under a variety of in vitro conditions. Furthermore, ␣-synuclein polymerization is temperature-, concentration-, and time-dependent, and the protein subunits are topographically organized within these polymers as reflected by the paucity of immunoreactivity for antibodies to ␣-synuclein epitopes at both ends of the polypeptide relative to those in the central region. Although a recent report demonstrated that wild-type ␣-synuclein forms Thioflavin-S-reactive aggregates and filaments, especially at elevated temperature (28), our studies substantially extend these preliminary observations by comparing the morphology of filaments formed from WT and mutant ␣-synucleins as well as the topography of ␣-synuclein in these filaments. A possible model consistent with these observations is that ␣-synuclein exists in numerous conformational states in aqueous solution, but when molecules with conformations compatible with dimerization interact, they may stabilize each other in this dimerized conformation, and these dimers may then serve as seeds for polymerization.
The findings presented here and in previous reports suggest that ␣-synuclein is the major building block of the filaments that form LBs. First, anti-␣-synuclein antibodies stain LBs more intensely and consistently than any other antibodies, including anti-ubiquitin antibodies (12,(15)(16)(17)19). Second, anti-␣-synuclein immunoreactivity also is abundant in pale bodies (19,24), which are believed to be precursor lesions of LBs. Third, neurofilaments, another filamentous component of LBs (29 -31), are detected only in a small percentage of pale bodies (32), and most anti-neurofilament antibodies only label a subset of LBs (30). However, at least one anti-neurofilament antibody that is specific for the mid-size neurofilament subunit (RMO32) labels nearly all cortical LBs (30). Fourth, the filaments generated in vitro from WT and mutant ␣-synucleins are morphologically similar to the LB filaments visualized in situ (30,31,33) as well as to synuclein filaments recovered from Sarkosyl-insoluble preparations of DLB brains (16), while normal neurofilaments have side arms (27,34) that are not seen in LB fibrils (31,33).
LB filament formation in vivo may result from ␣-synuclein accumulation, and this may be due to a reduction in the fast axonal transport (35) or overexpression of ␣-synuclein. Furthermore, once ␣-synuclein reaches a critical intracellular concentration, it may polymerize into filaments that aggregate into LBs or LB precursors which entrap other cytoplasmic components, and similar mechanisms may lead to the formation of Lewy neurites. Additionally, the reduced ability of A30P ␣-synuclein to bind vesicles (35) may play a mechanistic role in the onset/progression of some forms of familial PD, while the A53T mutation may be pathogenic because it increases the propensity of ␣-synuclein to polymerize into LB filaments (see Figs. 1, C-E, and 2). The accelerated formation of filaments as a consequence of the A53T mutation has also been observed in a recent publication by Conway et al. (42). Interestingly, Crowther et al. (43) have reported that carboxyl-terminally truncated ␣-synuclein may be more prone to form filaments than the full-length protein. Although the pathological implications of the latter finding is still unclear, it is possible that aberrant proteolysis of ␣-synuclein may create a pool of these shorter products that may play a role in initiating ␣-synuclein filament assembly.
Although ␣-synuclein has been reported to be expressed predominantly in neurons, it also is a major component of GCIs Filaments were generated as described in the legend to Fig. 3. WT and mutant ␣-synuclein filaments were labeled with various anti-␣synuclein antibodies followed by either a goat anti-mouse IgG antibody or a goat anti-rabbit IgG antibody conjugated to 10 nm gold (Electron Microscopy Sciences) as described in the text. Micrograghs A, C, and D illustrate WT ␣-synuclein labeled with antibodies SNL 4, Syn 202, and Syn 204, respectively. B and F depict A30P ␣-synuclein labeled with antibodies SNL-1 and Syn 208, respectively. E shows A53T ␣-synuclein decorated with Syn 204. A-F are the same magnification, and the scale bar equals 100 nm.

␣-Synuclein Filament Assembly 7621
and NCIs in MSA brains (20 -24). Ultrastructurally, GCIs and NCIs have been reported to be composed of 15-40-nm filaments (23, 36 -41), but some GCIs appear to contain tubular filaments (38,41), although this issue remains controversial (36). Recently, the filaments in GCIs have been shown to be intensely labeled with anti-␣-synuclein antibodies at the ultrastructural level (15,23), and Sarkosyl-insoluble ␣-synuclein filaments with straight and twisted morphologies similar to our synthetic filaments also have been observed in extracts of MSA brains (21). Perhaps the tubular appearance of some GCI fibrils is due to the incorporation of other proteins into ␣-synuclein filaments or these tubular fibrils may reflect the presence of microtubules trapped within bundles of ␣-synuclein filaments. This uncertainty notwithstanding, we speculate that the anomalous expression, accumulation, and subsequent polymerization of ␣-synuclein lead to the formation of GCIs in MSA brain. However, future research will be needed to determine whether ␣-synuclein polymers serve as a proteinaceous trap and if ␣-synuclein co-assembles with other proteins previously co-localized in LBs and GCIs in situ. Finally, efforts to over express WT and mutant ␣-synucleins in vivo will help to elucidate the role of ␣-synuclein filaments in the pathogenesis of PD, DLB, MSA, and related ␣-synucleinopathies.